26     March 2000

Accounting for Renewable and Environmental Resources

 , ablueribbonpaneloftheNationalAcademyof ing of the interaction between the U.S. economy Sciences’ National Research Council completed a congression- and the environment? The panel’s investigation, ally mandated review of the work that the Bureau of Economic while based on data that are highly imprecise and Analysis (BEA) had published on integrated economic and en- in some cases speculative, suggests that the de- vironmental accounts. The panel’s final report commended velopment of the accounts proposed for Phase BEA for its initial work in producing a set of sound and ob- III would be likely to encompass the most sig- jective prototype accounts. The November 1999 issue of the nificant economy-environment interactions. This SURVEY OF CURRENT BUSINESS contained an article by William D. Nordhaus, the Chair of the Panel, that presented an overview observation is tempered by the realization that to of the major issues and findings and a reprint of chapter 5, date nothing approaching adequate comprehen- “Overall Appraisal of Environmental Accounting in the United sive environmental accounting for a country of the States.” Chapter 3, “Accounting for Subsoil Mineral Resources” complexity of the United States has yet been un- was reprinted in the February 2000 issue; chapter 4, “Accounting dertaken. For BEA or the federal government to for Renewable and Environmental Resources” is reprinted below. prepare a full set of environmental accounts would This article is reprinted with permission from Nature’s Num- require a substantial commitment. bers: Expanding the National Economic Accounts to Include This chapter provides a review of the issues the Environment. Copyright of the National Academy Press, involved in accounting for renewable and envi- Washington DC. This is a report of the National Research Coun- ronmental resources. It is not intended to be a cil, prepared by the Panel on Integrated Environmental and comprehensive review of work in this area. Rather, Economic Accounting and edited by William D. Nordhaus and it delineates the issues that are involved in envi- Edward C. Kokkenlenberg. ronmental accounting and presents two important   chapter reviewed issues involved specific examples that illustrate these issues. The T in extending the national accounts to include first section reviews BEA’s efforts in environmental subsoil assets. This chapter focuses on two other accounting to date. Next, we analyze how stocks aspects of environmental accounting: renewable and flows of residuals from human activities relate and environmental resources. BEA has proposed to natural sources of residuals, natural resource covering these two categories of resources in fu- assets, stocks, flows, and economic activity. The ture work on integrated accounting. As discussed third section examines issues involved in account- in Chapter 1, Phase II of that work would focus ing for renewable and environmental resources. on different classes of land (e.g., agriculture, for- The chapter then turns to general issues associated est, and recreation land), on timber, on fisheries, with the physical data requirements of environ- and on agricultural assets such as grain stocks and mental accounting and with valuation. We next livestock. Phase III would address environmental investigate in greater detail the cases of forests and resources, including, for example, air, uncultivated air quality to illustrate how augmented accounting biological resources, and water. might actually be done. The chapter ends with the The general principles set forth in Chapter 2 in- panel’s conclusions and recommendations in the dicate that increasingly severe obstacles are likely area of accounting for renewable and environmen- to arise as the national accounts move further from tal resources. Appendix B identifies potentially the boundaries of the market economy. The dis- useful sources of data for developing supplemental cussion in this chapter confirms the premise that accounts identified by the panel in the course of its BEA’s Phase III raises the most difficult concep- investigation. tual, methodological, and data problems. This finding presents a dilemma that must be faced in BEA EFFORTS TO DATE IN expanding the accounts: Should follow-on efforts ACCOUNTING FOR RENEWABLE AND focus on those resources that can be most eas- ENVIRONMENTAL RESOURCES ily included given existing data and methods, or shouldBEAfocusonincludingthoseresourcesthat This section reviews BEA’s initial design for its would have the largest impact on our understand- supplemental accounts for natural-resource and     March 2000 27 • environmental assets. A more complete evalua- the purview of the environmental accounts ex- tion of BEA’s efforts on forests is included later tend? Table 4–1 shows BEA’s tentative decisions in the chapter. As discussed in Chapter 2, a on how it proposed to structure its supplemen- critical issue involved in the development of aug- tal accounts (BEA’s Integrated Environmental and mented accounts is setting the boundary. How Economic Satellite Accounts [IEESA] from Bureau far from the boundary of the marketplace should of Economic Analysis, 1994a: Table 1). Phase II

TABLE 4–1 IEESA Asset Account, 1987 [Billions of dollars] This table can serve as an inventory of the estimates available for the IEESA’s. In decreasing order of quality, the estimates that have been filled in are as follows: For made assets, estimates of reproduc- ible tangible stock and inventories, from BEA’s national income and product accounts or based on them, and pollution abatement stock, from BEA estimates (rows 1–21); for subsoil assets, the highs and lows of the range based on alternative valuation methods, from the companion article (rows 36–41); and best available, or rough-order-of-magnitude, estimates for some developed natural assets (selected rows 23–35 and 42–47) and some environmental assets (selected rows 48–55) prepared by BEA. The ‘‘n.a.’’—not available—entries represent a research agenda.

Change Opening Stocks Depreciaton, Revaluation Closing Total, Net Depletion, Capital and Other Stocks (3+4+5) Degradation Formation Changes (1+2)

(1) (2) (3) (4) (5) (6) Row

PRODUCED ASSETS Made assets ...... 1 11,565.9 667.4 –607.9 905.8 369.4 12,233.3 Fixed assets ...... 2 10,535.2 608.2 –607.9 875.8 340.2 11,143.4 Residential structures ...... 3 4,001.6 318.1 –109.8 230.5 197.4 4,319.7 Fixed nonresidential structures and equipment ...... 4 6,533.6 290.1 –498.1 645.3 142.9 6,823.7 Natural resource related ...... 5 503.7 23.1 –19.2 30.3 12.0 526.8 Environmental management ...... 6 241.3 8.4 –7.0 10.6 4.7 249.6 Conservation and development ...... 7 152.7 3.6 –4.4 5.3 2.7 156.4 Water supply facilities ...... 8 88.5 4.8 –2.5 5.3 2.0 93.3 Pollution abatement ...... 9 262.4 14.7 –12.2 19.7 7.3 277.1 Sanitary services ...... 10 172.9 12.8 –5.6 13.7 4.8 185.8 Air pollution abatement and control ...... 11 45.3 .6 –4.1 3.5 1.3 45.9 Water pollution abatement and control ...... 12 44.2 1.3 –2.5 2.6 1.2 45.5 Other ...... 13 6,029.9 267.0 –478.9 615.0 130.9 6,296.9 Inventories ...... 14 1,030.7 59.3 ...... 30.1 29.2 1,090.0 Government ...... 15 184.9 6.8 ...... 2.9 3.8 191.7 Nonfarm ...... 16 797.3 62.4 ...... 32.7 29.7 859.7 Farm (harvested crops, and livestock other than cattle and calves) ...... 17 48.5 –9.9 ...... –5.5 –4.4 38.6 Corn ...... 18 10.2 .3 ...... –1.1 1.4 10.5 Soybeans ...... 19 5.0 –.1 ...... –1.0 .9 4.9 All wheat ...... 20 2.6 0.0 ...... –.2 .2 2.6 Other ...... 21 30.7 –10.1 ...... –3.2 –6.9 20.6 Developed natural assets ...... 22 n.a. n.a. n.a. n.a. n.a. n.a. Cultivated biological resources ...... 23 n.a. n.a. n.a. n.a. n.a. n.a. Cultivated fixed natural growth assets ...... 24 n.a. n.a. n.a. n.a. n.a. n.a. Livestock for breeding, dairy, draught, etc ...... 25 n.a. n.a. n.a. n.a. n.a. n.a. Cattle ...... 26 12.9 2.0 n.a. –.3 2.3 14.9 Fish stock ...... 27 n.a. n.a. n.a. n.a. n.a. n.a. Vineyards, orchards ...... 28 2.0 .2 n.a. 0.0 .2 2.2 Trees on timberland ...... 29 288.8 47.0 –6.9 9.0 44.9 335.7 Work-in-progress on natural growth products ...... 30 n.a. n.a...... n.a. n.a. n.a. Livestock raised for slaughter ...... 31 n.a. n.a...... n.a. n.a. n.a. Cattle ...... 32 24.1 7.5 ...... 0.0 7.5 31.6 Fish stock ...... 33 n.a. n.a...... n.a. n.a. n.a. Calves ...... 34 5.0 .9 ...... –.5 1.4 5.9 Crops and other produced plants, not yet harvested ...... 35 1.8 .3 ...... 1 .2 2.1 Proved subsoil assets ...... 36 270.0 - 1,066.9 57.8 - 116.6 –16.7 - 61.6 16.6 - 64.6 58.0 - –119.6 299.4 - 950.3 Oil (including natural gas liquids) ...... 37 58.2 - 325.9 –22.5 - 84.7 –5.1 - –30.6 5.8 - 34.2 –23.1 - –88.3 35.7 - 241.2 Gas (including natural gas liquids) ...... 38 42.7 - 259.3 6.6 - 57.2 –5.6 - –20.3 4.1 - 14.9 8.1 - –51.8 49.4 - 202.2 Coal ...... 39 140.7 - 207.7 2.2 - 3.4 –5.4 - –7.6 4.4 - 6.3 3.2 - –2.1 143.0 - 204.2 Metals ...... 40 (*) - 215.3 67.2 - –29.5 –.2 - –2.2 2.2 - 9.2 65.2 - 22.5 38.5 - 244.8 Other minerals ...... 41 28.4 - 58.7 4.3 - .8 –.4 - –.9 .1 - .0 4.6 - .1 32.8 - 57.9 Developed land ...... 42 n.a. n.a. n.a. n.a. n.a. n.a. Land underlying structures (private) ...... 43 4,053.3 253.0 n.a. n.a. n.a. 4,306.3 Agricultural land (excluding vineyards, orchards) ...... 44 441.3 42.4 n.a. –2.8 45.2 483.7 Soil ...... 45 n.a. n.a. –.5 n.a. n.a. n.a. Recreational land and water (public) ...... 46 n.a. n.a. –.9 .9 n.a. n.a. Forests and other wooded land ...... 47 285.8 28.8 n.a. –.6 29.4 314.6 NONPRODUCED/ENVIRONMENTAL ASSETS Uncultivated biological resources ...... 48 n.a. n.a. n.a. n.a. n.a. n.a. Wild fish ...... 49 n.a. n.a. n.a. n.a. n.a. n.a. Timber and other plants and cultivated forests ...... 50 n.a. n.a. n.a. n.a. n.a. n.a. Other uncultivated biological resources ...... 51 n.a. n.a. n.a. n.a. n.a. n.a. Unproved subsoil assets ...... 52 n.a. n.a. n.a. n.a. n.a. n.a. Undeveloped land ...... 53 n.a. n.a. –19.9 19.9 n.a. n.a. Water (economic effects of changes in stock) ...... 54 ...... n.a. –38.7 38.7 n.a...... Air (economic effects of changes in stock) ...... 55 ...... n.a. –27.1 27.1 n.a......

n.a. = Not available NOTE: Leaders (...) indicate an entry is not applicable. * The calculated value of the entry was negative. Source: Bureau of Economic Analysis (1994a) SURVEY OF CURRENT BUSINESS, April 1994. The table has been slightly simplified for this report. 28 March 2000     •

of BEA’s development of supplemental tables fo- Other Sectors of the Proposed Accounts cused on assets listed in rows 22–35 and 42–47 of Table 4–1, while Phase III considers rows 48–55. As reported by BEA, the quality of actual entries Because BEA has not completed Phases II and III, in published supplemental accounts for Phase II actual decisions on what will be included have yet and III assets ranges from relatively good to con- to be made. Each of the following sections of this ceptually defective.1 For Phase II assets, estimates chapter considers an element of how to draw the within the category “developed land” are described line. While an ideal set of accounts would con- as “of uneven quality” (p. 45). According to tain “everything,” this chapter examines practical BEA, agricultural land values are “relatively good issues that arise in constructing actual accounts and are based on U.S. Department of Agriculture based on available data and tools. As will be seen, estimates of farm real estate values less BEA’s esti- the practical is likely to fall far short of the ideal. mates for the value of structures” (p. 45). BEA has not attempted to estimate the value of recreational land, but has entered federal maintenance and re- pair expenditures as an investment (see Table 4–1) Pollution Abatement and Control and “assumed that these expenditures exactly off- Expenditures set the degradation/depletion of recreational land” (p. 45). BEA indicates that this assumption is One particular entry in the environmen- made only for purposes of illustration and is “not tal accounts—pollution abatement and control to imply any judgment about the true value of expenditures—has been the subject of detailed in- degradation/depletion” (p. 45). A more detailed vestigation by BEA for many years. These items are discussion of BEA estimates for timber and land in shown for 1987 in rows5–12 of Table 4–1.TheBu- forestsispresentedlaterinthischapter. reau of the Census began collecting these data and For Phase III assets, BEA has entered “n.a.” BEA reporting them in 1972 (with some breaks in for most of the items, indicating that these esti- theseries); theseeffortsweresuspended in 1995 be- mates have not yet been developed. Entries for cause of budget cuts. Reporting of these costs does investment in and degradation of water, air, and not extend the accounts, but rather reorganizes the undeveloped land are included, however. As in existing accounts to provide a better indication of the case of developed recreational land, BEA has the interaction between the environment and the assumed that maintenance exactly offsets degrada- economy. tion, noting that this assumption provides entries that “are simply place markers” (p. 46). In the The limitations of these data are well recognized panel’s view, the use of maintenance expenditures and were discussed in Chapter 2. Many of the costs as degradation costs is highly misleading, and this included in the data overstate the cost of pollu- procedure should not be followed in the future. tion control, while other pollution-reducing costs Entering “n.a.” would be more accurate. The panel are omitted because they involve changes in pro- notes, however, that these estimates do not neces- cesses. There is also controversy about the extent sarily reflect BEA’s planned approaches, but were to which stringent pollution control regulations included by BEA to show the current state of data may have a chilling effect on innovation and tech- and research. nological change. Finally, little thought has been Regarding future plans, the United Nations Sys- given to the appropriate treatment of purchases tem of Integrated Environmental and Economic of emission permits, which are likely to become Accounting (SEEA) “does not recommend that the a more important feature of environmental reg- stock of air—which is truly a global common—or ulation in the future. Despite their limitations, water be valued; instead it recommends that valu- however, data on pollution abatement are likely ation be limited to changes in these assets—their to be among the most precise of the data in the degradation and investments in their restoration” environmental accounts, and they have been ex- (p. 46). It should be emphasized that the entries tremely useful for understanding trends and levels for environmental assets in Table 4–1 are highly in control costs and for examining how environ- oversimplified. Some components of air quality, mental programs have affected productivity. The such as greenhouse gases and stratospheric ozone, panel finds that the data on pollution abatement are truly global assets and services; others, such as expenditures are valuable and, as noted in the final reductions in urban smog, are local and regional section of this chapter, recommends that funds be

provided to improve the design and recommence 1. All quotations in this section are from the Bureau of Economic Analysis collecting these data. (1994a).     March 2000 29 • public goods. Additional dimensions that need to POLLUTANT EMISSIONS AND THEIR be incorporated are relations to external events, RELATION TO STOCKS, FLOWS, AND spatial resolution, and nonlinearities in damages. ECONOMIC ACTIVITY The discussion of air quality later in this chapter illustrates its multiple dimensions. Similarly, wa- Before constructing environmental accounts, it is ter quality and quantity, undeveloped land, and necessary to determine the interactions between uncultivated biological resources are composites natural resources and the environment and eco- of many different assets and quality characteristics nomic activity. It is essential to understand the that provide multiple goods and services. key physical flows and stocks and how they affect BEA’s efforts have focused on the asset accounts. humans and economic activities and values. A A preliminary table for a production account with- complete accounting requires detailed knowledge out entries is included in BEA’s report on its of the physical properties of resources and pollu- development of the IEESA (Bureau of Economic tants as described in fate, transport, and impact Analysis, 1994a, 1994b). Production of market ordamagemodels,aswellastheserviceflowsto goods and services from these natural assets—e.g., market and nonmarket sectors. timber, agricultural crops, fish—is already in the Figure 4–1 illustrates key relationships among core production accounts. Greater attention is emissions, stocks of pollutants, natural-resource needed to identifying, measuring, and valuing the assets, and economic activities in different sectors. specific types of nonmarket goods and services As the figure shows, economic activities produce produced by these assets. a variety of uninternalized emissions and resid- 30 March 2000     •

uals that find their way into the environment. syndrome”—the fact that the same substance may Manyofthepollutantsofconcernareresidualsthat be beneficial or harmful depending on where it also have natural sources—sulfur, carbon dioxide, is, how much of it there is, the time and dura- carbon monoxide, nitrogen compounds—and are tion of exposure, and what organism is absorbing emitted during volcanic eruptions, produced by it. Virtually every substance on earth, from wa- forests and wetlands, or released from wildfires. ter to plutonium, can be an economic good or an Other residuals of concern—such as chlorofluo- economic weed depending on the circumstances. rocarbons (CFCs) and many pesticides used in One of the most important difficulties is that the agriculture—areanthropogenic and have no natu- physical measurements used are often inaccurate ral sources. In terms of effects on human activities, indicators of actual human exposures. Average the sources of the residuals are not important. emissions of the precursor pollutant, average con- What may be important is that human activities centrations over the year, or concentration data for have increased the levels occurring in the environ- limited sitesaregenerallynot representativeofcon- ment, concentrated them to a degree that makes centrations to which the population is exposed and them dangerous, or relocated them to areas where may be a misleading basis for developing damage people or economic activities are exposed to them estimates. For example, tropospheric ozone forms at high levels. mainly in warm weather. Thus total annual hydro- Whether from natural sources or human activi- carbon emissions, the precursor to tropospheric ties, environmental variables can affect economic ozone, are a poor indicator of potential levels of well-being in three general ways, as illustrated in tropospheric ozone. Tropospheric ozone levels Figure 4–1: (1) direct effects on consumption or also vary significantly over the distance of a few city income of households, industry, and government; blocks. One of the major challenges both for better (2) accumulation in the environment of stocks of environmental policy and for the construction of residuals that then affect economic activities or environmental accounts is to obtain better meas- economic assets; and (3) effects on the service ures of direct human exposure to the important flows of economic assets (capital stock, natural re- harmful substances among a representative sample sources, or human resources), such as recreation, of people. clean air to breathe, and navigable river channels free of sedimentary deposits. Accumulation of Stocks

Direct Effects Many environmental problems result from the ac- cumulation of residuals. These substances include Environmental variables affect human and natural most radiatively active trace gases, which remain systems directly. Urban smog, whose concentra- in the atmosphere for decades or centuries, and tions change daily or even hourly, is an obvious many radioactive materials, which have half-lives example. Sulfate and nitrate aerosols, pollutants of decades or centuries. Similarly, recovery from contributing to acid precipitation, remain in the stratospheric ozone depletion is a process requir- atmosphere for a matter of days. These pollutants ing years or decades. and agricultural chemicals have short-term health effects, reduce visibility, often migrate very slowly through soils, contam- interfere with recreational activities, affect crop inating drinking water only after several years or growth, and present their own set of problems for decades. accounting. In many cases, the substances emitted Environmental accounting therefore needs to are precursor emissions; that is, they react chem- develop and include appropriate methods to ac- ically in the atmosphere with other substances to count for those persistent pollutants, such as heavy form the substance that is ultimately damaging to metals that accumulate in the environment and humans or ecosystems. There are also complex last for many years. Each year’s emissions or nonlinearities because the formation of the dam- production of residuals adds to the stock in the aging substance depends on the level of precursor environment, and it is necessary to understand the emissions, weather conditions, and the presence of processes by which these stocks decay or dissipate. other substances with which the precursor emis- In some cases (as with radioactive substances), sionsreact. Alloftheseprocessesvaryonanhourly, those processes are easily understood, while in daily, and seasonal basis. Emissions, concentra- other cases (such as subsoil toxins or the carbon cy- tions, and impacts of damaging substances also cle), understanding the processes poses enormous vary spatially, and there may be important thresh- scientific challenges. In the economic accounts, the old effects as well. Above all, there is the “weed stock-flow dynamics are similar to those of gross     March 2000 31 • investment and depreciation of capital. While final demand. Current-year activities would in- there is a conceptual similarity, however, there is clude production of residuals, just as traditional no readily observable market price for these stock economic accounts include production accounts. changes. Hence, valuation of a change in stock A complete set of accounts would incorporate requires estimating the valueof the impact of addi- flows of residuals from abroad, similar to im- tions over the lifetime of the stock, accounting for ports of goods and services. It would also be dissipation, and appropriately discounting future necessary to calculate the “price”—negative or effects. It should also be recognized that, with a positive—indicating whether theeffect wasadverse few exceptions, the stocks are extremely heteroge- or beneficial. The accounting for current-year neous, so that measuring a simple “environmental activities would include final uses of residuals, capital stock” is likely to be extremely difficult. identifying effects on final consumption, flows abroad, and contributions to capital stocks, just Effects on Economic Assets as traditional accounting frameworks identify final consumption of goods and services, exports, and Both short-lived and long-lived residuals can affect gross capital accumulation. economic activity over a number of years through their effects on other economic assets, in particu- Accounting for Capital Assets lar produced capital goods such as buildings and equipment. For example, acid precipitation can It is important to measure the volumes and val- cause deterioration of buildings. Accumulated ues of the nation’s natural assets for many reasons. greenhouse gases can result in coastal flooding and One purpose is simply to determine general trends. higher storm surges, thereby adversely affecting the Another, illustrated in Table 4–1, is to determine value of existing coastal structures. Pollutants such the relative magnitudes of different assets. A fur- as lead can cause long-lasting health consequences, ther reason arises in the context of sustainable impacts on intellectual functions, and premature economic growth. As discussed in Chapter 2, one death. can calculate measures of sustainable income if one corrects conventional measures of national income ISSUES INVOLVED IN ACCOUNTING by including the value of the change in the stocks FOR RENEWABLE AND of natural and other assets. ENVIRONMENTAL RESOURCES For all of these reasons, we would ideally like to have measures of the value and volume of the The previous section addressed the major ways nation’s natural assets; thus we must include meas- in which natural resources and the environment ures not only of “made assets,” such as houses interact with economic activity. Depending on and computers, but also renewable resources, such the intended uses of the data, there are differ- as timber or the fertility of land, and nonrenew- ent approaches to structuring environmental and able assets, such as oil and mineral resources. natural-resource accounts. The most complete It is important to know whether the economy accounting structure would treat all the relation- is generating an ever-growing stock of damaging ships in Figure 4–1. However, constructing such environmental residuals that will pose a large eco- a complete set of accounts is infeasible today, and nomic burden on future generations. We want to governments must choose areas for investigation know whether the economic value of investments strategically in accordance with their national eco- in tangible, human, and technological capital is nomic and environmental goals and interests. This more than offsetting whatever depletion of natural section delineates some possible approaches to ac- assets is occurring. counting for natural and environmental resources There is a close connection between the pro- and activities. duction accounts and the asset accounts (see Chapter 2). As noted above, measures of compre- Production and Income Accounts hensive income or of sustainable income include not only current consumption flows, but also the A complete set of production accounts would iden- value of the change in the stocks of assets. Hence tify all the cross-relationships among industry, augmented accounting requires careful and accu- household, government, and natural sources of rate measurement of both assets and consumption emissions or residuals, as well as the nonmarketed flows. Such measurement is currently undertaken current account input services provided by na- within the boundary of the marketplace, but aug- ture and the productive contribution of nature to mented accounting would require extending that 32 March 2000     •

boundary for both assets and consumption in a be similar to redefining an automobile as a com- consistent manner. The conceptual basis for as- bination of transportation mode, public-health set valuation in environmental accounts parallels menace, and status symbol. Under this approach, closely that in the conventional accounts. De- an asset is valued in terms of the sum of the val- pletion and degradation of natural resources is ues of its various characteristics. In this view, conceptually similar to depreciation of produced there is little point in trying to analyze the to- capital assets. Stocks of residuals can decay or tal value of holistic assets such as land or air or dissipate, a process that is again conceptually sim- climate; rather, one undertakes the more modest ilar to depreciation of produced assets. Natural task of looking at the different functions involved.2 growthofbiological resources, rechargeofground- BEA’s treatment of soil erosion is consistent with water resources, and accumulation of residuals are this approach; agricultural land is treated as the conceptually similar to gross capital formation or asset and the soil depth and organic-matter con- investment. Net accumulation of assets is equal to tent as characteristics of the land. Other aspects thevalueofthechangein stocks. Manyoftheissues of land quality—local climate or ambient level of involved in constructing chain indexes of values pollution—can be considered in a similar manner. and volumes translate directly into measurement Identification of the economic effects of erosion on of resource and environmental stocks. the value of land makes the resource link explicit. However, some special conceptual difficul- Thus, a potentially useful alternative to consid- ties arise in measuring stocks of natural assets. ering the holistic value of assets is to consider Natural-resource assets (like a physical plant or how changes in air quality affect the value of agri- piece of equipment) are complex systems of com- cultural land, forests, residential property, and ponent parts that have value because of the way human capital. Thus, fundamental nonhuman as- they work together. Since produced capital assets sets might include forests, lakes, rivers, estuaries, are generally purchased or constructed as modules, coastal regions, wetlands, farmland, and residen- they can be valued on the basis of their own market tial property. This approach has two further prices, rather than their synergistic contribution attractive features: it allows better integration with to output. To take an analogy, a baseball player’s existing accounts, since some of these assets (such contribution to the team is a complex function not as residential property and forests) have an exten- only of hitting, pitching, and fielding, but also of sive existing database; and it allows incremental temperament, teamwork, and verbal abilities; from development of a set of valuations, building upon an accounting perspective, however, the economic those in the market sector. contribution is simply wages and other compen- sation. For environmental assets, determining the Practical Choices in Expanding the value will become difficult when the effort extends Accounting Framework beyond the market boundary. Consider a forest. How can the value of the stumpage in the forest be A complete accounting system including interac- separated from the forest’s contribution to erosion tions in the production and asset accounts would control, air quality, and biodiversity? be a significant undertaking. Deciding on the scale Even when markets produce evidence of the of augmented accounting and the next steps to be value of a bundle of assets—the composite value of taken will require considerable strategic thought. soils, timber, nearness to water, and recreation—it One question is whether the accounts will be may be difficult to separate out the values of the dif- used for scorekeeping or for management (see the ferent components without applying complicated discussion in Chapter 2. statistical procedures. Sometimes, the separation Scorekeeping, which involves developing a better is misleading, as when the value of the components measure of the performance of the economy over depends on their being together. An assembled time, is one perspective. It addresses the questions bicycle is different from a pile of parts; similarly, of trends in the values of environmental assets and forests, lakes, rivers, farmland, and coastal es- whether current consumption is sustainable. If tuaries are valuable because of the way they are scorekeeping of this type is the purpose of sup- assembled. plemental environmental accounts, it will simplify One possible way of avoiding this difficulty is to the enterprise because there will be no need to redefine assets in terms of particular functions or consider intermediate interactions between pro- characteristics, an approach similar to that taken duction sectors. Tracing where pollutants were

in hedonic valuation, whereby goods are viewed as 2. Watershed valuation is an example of a holistic approach (see Anderson packages of characteristics. This approach would and Rockel [1991] and Green et al. [1994] as examples).     March 2000 33 • produced and how they affect intermediate prod- processes come into play. Consider the problems uct is unnecessary as long as one can measure final involved in accounting for a simple loaf of bread. consumption and changes in assets. For example, Doing so would require measuring and valuing a dying forest is a deteriorating asset; whether the a wide variety of flows of water, fertilizer, pesti- deterioration is caused by acid precipitation, tro- cides, labor, climate, and capital inputs that go into pospheric ozone, or pest infestation is secondary producing the wheat; the fuels, transport vehicles, from a scorekeeping perspective. What is im- emissions, weather-related delays, induced con- portant is to measure the deterioration accurately. gestion, or floods involved in transportation; the Similarly, the overall health and skills of human molds, spores, and miscellaneousrodentsand their populations is a central issue in measuring whether droppings that invade the storage silos; the com- theeconomyascurrentlystructured isleading to an plex combination of human skills, equipment, and increase or decrease in the stock of human capital. structures that go into milling the wheat; the en- Why the change is occurring—whether because of trepreneurship of the baker and the software in the changes in health care or education expenditures computer-operated baking and slicing machinery; or reductions in blood lead—is secondary to the the complex chemistry and regulatory environ- measurement issue. Overall scorekeeping would ment involved in the wrapping materials; and the note the substantial improvements in the health evolving ecology of the distribution network. Be- status of Americans over this century rather than hind each of these elements, in addition, is the decreases in particular ailments. complex general equilibrium of the marketplace, The second broad perspective on the function which determines the selection of production pro- of environmental accounts is that of environmen- cesses by prices, taxes, and locations, along with tal management. This perspective focuses on the further complexity of needing to unravel the the sources, transportation, and ultimate disposal input-output structure of the inputs into each of of residual pollutants, particularly their contri- the steps just described. butions to outcomes of economic and ecological It appears unlikely that anyone would try, and consequence. Knowing to what extent partic- safe to conclude that no one could succeed in, ular emissions of residuals come from utilities, describing the physical flows involved in this lit- automobiles, or volcanic eruptions is critical to de- tle loaf of bread. Fortunately, however, economic veloping strategies for control. If human sources accounting does not attempt such a Herculean are dwarfed by natural sources, for example, efforts task. Rather, the national accounts measure all to control human sources may be futile. Simi- these activities by the common measuring rod of larly, knowing that life expectancies have increased dollars. Although the dollar flows are routinely dramatically is not very useful to understanding broken down into different stages—wheat, trans- whether there are benefits to tightening controls portation, milling, baking, and distribution—one on small particles or ozone. Improvements in could never hope to describe the flows physically health care, occupational safety, and traffic safety and then attach dollar values to each physical stage. may result in increasing life spans and health status Yetthisisjustwhatwouldberequiredforafull more than pollutants are shortening life span—but and detailed set of environmental accounts. The reducing pollution further could extend lives fur- above comparison may give some sense of why ther. Thus, if the supplemental accounts are meant accounting for environmental flows outside the to support environmental management decisions, marketplace is such a daunting task. knowing the sources of pollutants and the specific causes of changes in asset quality are essential. PHYSICAL DATA REQUIREMENTS: GENERAL ISSUES Analogy with Economic Accounts Some of the analytical questions involved in envi- The discussion in this section has emphasized the ronmental accounting have been analyzed in the complexity involved in constructing environmen- previous section. To construct actual accounts tal accounts. It is useful to compare environmental requires both obtaining accurate physical data with conventional economic accounting. A lit- (discussed in this section) and valuing the flows tle reflection suggests that economic activity has a (discussed in the next section). similar, almost fractal complexity when one looks Accurate data on physical flows and stocks are a under the surface. It would be just as difficult to prerequisite for developing any accounting system measure the physical flows in economic life as in and are the focus of national accounting systems environmental life, and indeed many of the same under development in several European nations. 34 March 2000     •

In some areas, ample physical data are available as a In addition to understanding the dose-response by-product of regulatory monitoring and resource relationship, national accounting requires regular, management systems. Appendix B lists a number statistically valid monitoring of the relevant pop- of databases identified by the panel that may be of ulations and the doses they are receiving. A basic use in further work on supplemental accounts. limitation of much of the data currently collected Three concerns are fundamental to understand- is that ambient concentration levels in areas where ing data and measurement requirements for the individuals, crops, forests, or other relevant en- development of environmental accounts: (1) the tities actually reside are poorly measured. Most dose-response relationship, (2) measurement of measurements occur at sites of convenience rather actual doses experienced, and (3) the fate and than sites of relevance. Air pollution monitors are transport of residuals in the environment. The often placed with other monitoring devices where first, the dose-response relationship, is the physical airplanes congregate rather than where people live. relationship between the concentration of or expo- sure to an environmental change and the response Afull accountofeconomic-environmental inter- of the subject experiencing the dose. The dose- actionsalso requirestracking the fateandtransport response relationship is applied to many different relationship, or the connection between the emis- situations, for example, the response of trees and sion ofa particular pollutantorpollutantprecursor crops to chemicals such as carbon dioxide, tropo- at one time and geographic point and the level, spheric ozone, or acid deposition and the response time, and location of the pollutant at the point of humans to pollutants such as lead, particulate whereit affects an economic asset or activity. These matter, or radiation. relationships are generally highly complex and variable. For air pollutants, wind direction and Dose-response relationships are often difficult to speed, temperature, cloudiness, and precipitation determine because they may be affected by com- all affect how a pollutant is dispersed or concen- plex interactions and intervening factors. For trates. Precursor pollutants sometimes do not example, there are extensive medical data on causes create damage themselves, but react chemically in of death and, less universally, illness. To deter- the atmosphere to create other agents that are dam- mine impacts of environmental changes on human aging. Acid precipitation and tropospheric ozone or natural ecosystems requires separating out the are examples. The formation of these pollutants different causes of premature death or illness. In depends on the presence of other agents that may some areas, such as the impact of tobacco or lead, limit, speed, or slow the process. Monitoring of the relationships are relatively well established; in emissions, concentrations, exposures, and conse- other areas, such as the impact of particulate mat- quences would provide the physical foundation for ter or ozone, much uncertainty persists. For many a complete set of environmental accounts, and is of these relationships, average exposure over the also a critical part of environmental management. year is rarely the relevant measure. Damage may be related to extreme levels or to periods in which The goals of environmental accounting will dic- the subject is particularly sensitive to the agent; tate the assignment of priorities for improved acute effects may differ from chronic effects related data. Extensive data on the fate and transport to long-term, low-level exposure. of emissions and concentrations of pollutants are Resolving these uncertainties about dose- a lower priority if the goal is scorekeeping; even response relationships is important for policy dose-response relationships may be secondary to decisions, such as the level at which to set pri- more direct measurement of consumption flows or mary air-pollution standards. Resolution of these changes in important capital and environmental uncertainties would also allow construction of en- assets and human health status. If one is inter- vironmental accounts. The panel’s review of work ested primarily in measuring the sustainability of in this area indicates that the preparation of esti- economic activity, understanding the health status mates of the economic impacts of air pollution is of human and natural systems is more important feasible today, but there are enormous uncertain- than understanding why conditions have changed. ties at virtually every stage of the effort. While On the other hand, understanding these technical BEA or those preparing environmental accounts relationships is essential if environmental accounts would not necessarily be involved in preparing are to serve as a data set to support environmental dose-response estimates, the accountants will need management, in which the goals are to understand to work closely with public-health, agricultural, the severity and causes of environmental prob- forestry, and ecological experts to use the best lems, along with remedies needed to mitigate those information available. problems.     March 2000 35 •

VALUATION: GENERAL ISSUES been developed. Table 4–2 indicates the poten- tial and actual uses of various valuation methods Once appropriate physical data have been de- for many environmental problems, including the veloped, the next step in developing integrated dose-response method discussed above. These accounts is to value changes in the physical meas- methods have been developed over a number of ures. Physical data alone are often interesting and years and have been applied to many specific useful for policy making, and improvements in, problems.3 physical environmental data could enhance policy- The dose-response method,as a valuation method making efforts. Indeed, most countries have not in and of itself, is directed toward converting expo- gone beyond developing physical measures and sure to a specified dose of a substance, from which indicators because of the difficulties involved in is calculated a physical response for which a di- valuing nonmarket goods. Without valuation, rect market price can be observed. For example, however, physical data alone have serious limi- exposure to ozone or particulate matter results in tations for both scorekeeping and environmental wheat-yield loss or lost work-days due to respira- management. Aggregate physical measures, such tory illness; using the market price of wheat or of as areas of agricultural land, forest, or wetlands or labor, an estimate of economic value can be made. tons of sulfur, toxic wastes, or particulate emis- The valuation techniques in this approach are con- sions, provide incomplete second column evidence sistent with prices used in the economic accounts. on the effects of these chemicals on economic well- Incomparability or additional uncertainties are in- being or economic sustainability over time. For troduced only through imputation of output by example, losing 1000 acres of prime Florida Ever- use of the dose-response relationship, which con- glades would probably impose a greater economic verts the environmental effects into market-good and ecological loss than losing an equivalent area terms. of frozen wetlands in northern Alaska. Thus an ac- counting entry of “total wetland acres” lost would Travel-cost and hedonic methods also use behav- not be a useful measure. Furthermore, a sim- ior and observed market transactions as a basis for ple measure of wetland area would fail to capture estimating values, but the activities involve time improvements in quality that might occur as a re- use and expenditures on goods and services related sult, for example, of current efforts to restore the to use of the environmental or natural-resource Everglades as a fully functioning ecosystem. good, rather than on the resource itself. For ex- For many issues, it is necessary to weight the ample, a recreational site might be valued using physical measures by their importance. There are the travel-cost method by estimating the time and approaches to weighting physical quantities other out-of-pocket costs involved in reaching the site. than valuing all impacts in dollar terms; for ex- Hedonic methods use statistical techniques to ample, different environmental residuals can be explain variations in market prices based on the weighted by how they affect human mortality. bundle of characteristics of a good. This ap- However, such weights would be incomplete be- proach is currently used in the national accounts. cause they would exclude impacts on morbidity or Computers, for example, are considered bundles on the health of ecosystems. In economic account- of attributes such as speed, memory, and ran- ing, the “importance weights” are the economic dom access memory (RAM), and the value of the values, usually market prices. The advantage computer is a weighted sum of the values of its of using economic valuation is that comparisons attributes. can be made across very different environmental For resource and environment valuation pur- effects and with goods that are part of the mar- poses, hedonic methods are used to explain ket economy. While relying on economic values variations in land values that reflect natural- has many desirable features, there are a num- resource or environmental characteristics. Such ber of difficulties involved in usefully applying estimates are based on observed price differences nonmarket valuation studies and techniques to en- of land with different amenities or disamenities vironmental accounting, as discussed below (see such as noise, pollution, and crime. Hedonic wage also Chapter 2). studies—looking at the wage premiums of high- risk jobs—are currently the standard approach to Valuation Techniques estimating the value of workplace hazards; the re- sults are often used as estimates of the value of Markets provide the conventional valuation for market goods and services. A variety of meth- 3. See Smith (1993) and Braden and Kolstad (1991) for reviews of the ods for valuing nonmarket goods and services has theory and application of these methods. 36 March 2000     •

life-threatening effects due to such causes as air incurred. While widely used for environmental pollution or traffic accidents. valuation, CV is highly controversial because it of- ten fails elementary tests of consistency and scaling Contingent value (CV) methods are survey tech- and issubject to a widevarietyofpotential response niques that ask people directly what they would errors if not carefully constructed. pay for goods and services. Applications in the area of environment and natural resources include, The overriding problem with all these methods for example, asking individuals what they would is that they require voluminous data and statisti- be willing to pay to reduce smog, to increase visi- cal analysis and can hardly be used routinely for a bility in places such as the front range of Colorado, large number of products in constructing environ- and to clean up an oil spill in a coastal area. CV mental accounts. Where existing CV studies are methods differ from the other methods discussed used for environmental or natural-resource val- above in that there are no budget constraints or be- uation, they often employ valuation approaches havioral observations involved; the results reflect that are inappropriate for national accounts. For respondents’ estimates of the value of a hypotheti- example, many estimates used in environmental cal change, rather than a dollar or time cost actually management rely on average value (including con-

TABLE 4–2 Methods for Environmental Valuation

Techniques for estimation impacts Pollution Type of Effect Impact Hedonic Contingent Dose Hedonic Property Wages Travel Cost Valuation Response

Air pollution Conventional pollutants: Respiratory WLD L L X U U (total suspensed illness RAD particulate [TSP], Medical sulfur dioxide [SO2], suffering nitrous oxides [NOX]) Respiratory Death L and U U X X U illness Aesthetics Visual, U L X U X sensory Recreation Visits, especially to LXUUX forests Materials Maintenance/repair X X Poss Poss U Vegetation Crop losses L X X X U Water pollution Conventional pollutants Recreation Visit behavior L X U U X (e.g., biochemical (e.g., fishing, oxygen demand [BOD]) boating) Commercial Stock losses X X X X U fisheries Aesthetics Turbidity, odor, U X L U X unsightliness Ecosystem Habitat and X X X U U species loss Trace concentrations Drinking Illness, X X X Poss U water mortality Fisheries Stock losses X X X X U Toxic substances Air (benzene, Illness, WLD U X U U U polychlorinated mortality RAD Biphenyls [PCBs], Medical expenses pesticides) Pain and suffering Chemicals hazardous to land Aesthetics Unsightliness X X X U U Ecosystem Anxiety, ecosystem losses Radiation Illness, WLD Poss U X L U mortality RAD Lives lost Marine pollution Oil, radioactive Aesthetics Unsightliness U X U U U substances,sewage Swimming Visit behavior Illness Fish/livestock losses Noise Nuisance Annoyance U X X U L

U = Used technique; Poss = Not developed, but possible; X = Inapplicable technique; WLD Source: Adapted from Organization for Economic Cooperation and Development (1989), as = Work loss days; L = Very limited applications; RAD = Resource activity days. appearing in Costanza (1997).     March 2000 37 • sumer surplus), rather than the prices or marginal old or young, rich or poor, American scientist or values that are the convention in national income African farmer—one will benefit from the eradi- accounting.4 In a competitive economy, market cation whether one wants to or not. The example prices measure both the incremental value to the of smallpox eradication is a dramatic case of a economy of consuming another unit of the good public good. The economy is replete with activ- and the incremental cost to the economy of pro- ities, such as pollution abatement, new scientific ducing that unit. Therefore, prices are a useful knowledge, national defense, and zoning, that have benchmark for valuation. public-good characteristics.5 In one sense, the market value underestimates The distinction between public and private the total value of goods and services to consumers. goods is central for many nonmarket and environ- Because consumers pay the price of the last or mental commodities. In a perfectly competitive marginal unit for all units consumed, they enjoy market, the price of a marketed private good is the a surplus of total satisfaction over total cost. The marginal value of consumption to the consumer. term used for the extra utility consumers receive Similarly, while observed prices do not exist for over what they pay for a commodity is consumer nonmarket private goods, the marginal value of the surplus (see also Chapter 2). Consumer surplus consumption of such goods is conceptually equiv- introduces a complication in comparing market alent to a market price. The national accounts prices with nonmarket values. For goods without value food produced and consumed on farms,even markets, value is often measured by total willing- though it is not marketed, the same way food sold ness to pay for the good. Such values are not in the marketplace is valued. directly comparable to market prices because the values include the consumer surplus. In other Valuation ofpublic goodsisan especiallydifficult words, when nonmarket goods are valued accord- problem because their value to all consumers must ing to total willingness to pay, the value of those be reckoned with. For example, improvements in goods is overstated relative to the market value of air quality affect everyone. Conceptually, there- marketed goods. For example, travel costs can fore, one should value public goods by adding up provide the average value of a recreational serv- themarginal valuesofchangesto theentireaffected ice, but the marginal value of the resource for an population. Doing so poses severe measurement open-access beach or forest with no fee may be difficulties for two reasons. First, the “personal zero. This discussion illustrates the importance of prices” or marginal values of the public good are ensuring comparability in estimating values in the sure to vary across people—some may be signif- construction of nonmarket economic accounts. icantly affected and therefore place a high value on air quality, while others may be relatively indif- ferent. Second, determining the values of public Classes of Economic Goods goods is extremely difficult because people make few decisionsthat reveal their preferences in thisre- The valuation of environmental goods and serv- gard. People cannot choose how much defense or ices raises an issue that is largely overlooked in smallpox eradication they would like to consume; conventional accounting—thedistinction between these decisions are made collectively. Since peo- private and public goods. These deceptively com- ple cannot choose different levels of a public good, mon terms are used in a specialized sense here (see Samuelson, 1954, 1955). Private goods are ones that can be divided up and provided separately to 5. This discussion greatly simplifies the discussion of public goods. There different individuals, with no external benefits or are further distinctions among public goods that are central to many issues involved in environmental accounting, particularly as regards valuation meth- costs to others. An example is bread. Ten loaves ods. One such distinction is whether consumption is excludable; in the case of bread can be divided up in many ways among of global warming, for example, no coastal nation can exclude itself from the rising seas. Another distinction is between pure and congestible public goods. individuals, and what one person eats cannot be Congestible public goods are those whose consumption is neither completely rival nor nonrival; one person using a beach does not preclude others from do- eaten by others. Public goods, by contrast, are ones ing so, but most people find crowded beaches less enjoyable than deserted ones whose benefits are indivisibly spread among the (see Cornes and Sandler, 1986). Crowding of this sort means that even with open access, the marginal value of use of these sites is greater than zero. A final entire community, whether or not individuals de- distinction is between those goods whose use affects market activities or market sire to purchase them. An example is smallpox values and those that are completely independent of the market. Public goods without traces in markets are frequentlyreferred to as “nonuse values.” Nonuse eradication. It matters not at all whether one is values include values people derive from knowing that a species exists, natural wonders remain, or natural systems survive intact beyond any specific use to 4. Marginal costs and marginal values are central concepts in determining which they might be put (see Randall and Stoll, 1983). When Congress created economic efficiency. For example, knowing the marginal value of reductions Yellowstone National Park in 1872, for example, no member of Congress had in atmospheric lead is more useful to the policy maker than knowing the ever been there, and its value as a natural wonderland was largely a “nonuse average value of all reductions. Marginal cost and marginal value are defined value” imagined on the basis of photographs of William Henry Jackson and in Appendix D. drawings of Thomas Moran. 38 March 2000     •

there are no behavioral traces of their preferences not currently in the accounts, such as recreation or personal prices. services enjoyed by households; in this case, the For the above reasons, constructing environ- value that is attributable to the service is equal to mental accounts will necessarily be different for the value of household labor and capital services, private and public goods. For private goods, plus a service flow from a natural resource. particularly near-market goods that have close rel- Public goods that affect markets offer oppor- atives in the market economy, valuation appears tunities for using observations of actual market feasible and has a level of reliability that approaches transactions to generate valuation estimates. An that of the current national income accounts. Most example would be concessionaire activity within public goods, by contrast, present greater measure- a national park. The hedonic property and wage ment and conceptual problems. Table 4–3 shows techniques can be explored as a basis for develop- examples of each type of goods that have these ing valuesor imputing how changes in these public different characteristics. goods affect markets. There are some potentially sound ways to make the links between these pub- Strategies for Valuation lic goods and the market explicit in the accounts, but there is not yet a consensus on how to in- Near-market natural-resource and environmental clude them, and each provides a challenge for data goods (which are largely private goods) offer the development and estimation of values. most promise for valuation and inclusion in the Other classes of public goods, particularly those accounts. Often there are markets for comparable that are national or global in nature and do not goods that provide direct evidence of the value of leave behavioral traces of individual preferences, the nonmarketed goods or services. This approach are currently problematic for the national ac- is consistent with the use of market prices used counts. Most of these public goods, such as those elsewhere in the accounts and has precedent in involving nonuse values of natural-resource and the valuation of owner-occupied housing services. environmental assets, can be valued only with CV Thus, the methods for including these near-market methods. Some reviews have conveyed cautious goods have already been established. A potential approval for use of these methods in limited cir- source of error in using this approach is that the cumstances. For example, a panel convened by qualitymaydiffer for goodsor servicesproduced or the National Oceanic and Atmospheric Adminis- provided in the household and those produced in tration to review CV methods for use in federal the market. It would be appropriate to undertake a compensation decisions identified “a number of modest research program to investigate the adjust- stringent guidelines for the conduct of CV studies” ments necessary to make market and near-market that, when followed, allow “CV studies [to] convey activities comparable. useful information” (see Arrow et al., 1993:4610). Two basic types of near-market goods are of in- However, the accuracy of the values developed with terest. The first is the service flow from a natural these methods remains controversial among those resource. Here, as in the case of timber from forests in the economics profession (see Portney, 1994; or crops from farmland, the service flow is already Hanemann, 1994; Mitchell and Carson, 1989; and in the core accounts, and the returns to these assets Diamond and Hausman, 1994). appear as profits and/or returns to other assets, but As discussed above, the hypothetical nature of the accounting is incomplete because it omits the the valuation makes these methods quite differ- nonmarket activities. The second case is a good ent from other methods that are based on actual market transactions. For these reasons, while CV TABLE 4–3 Classes of Goods and Services is sometimes useful for other purposes, the panel has determined that it is currently of limited value Public (examples) Type of goods Private (examples) for environmental accounting. This means that, Related to Markets Independent of Markets for many important environmental assets, envi- Market Bread Knowledge and innova- None ronmental accounts will omit a portion of the Cars tions that are patented Restaurant meals and copyrighted value of the assets. That is, it appears to be fea- Housing rentals Pollutants with tradeable permits sible to work toward accounting for goods such Nonmarket Household prepared Air and water quality Passive or nonuse value as recreation activities associated with the Florida meals Climate (e.g., knowledge of the Everglades, Yellowstone National Park, and similar Leisure time Mosquito control existence of species, Television viewing unique national treas- sites. However, it is beyond the ability of current Groundwater for drinking ures such as Yellow- Rental values of owner- stone National Park) techniquesto providereliablemeasuresofthevalue used assets of the public-goods services provided by these as-     March 2000 39 • sets, even thoughwe maysuspect that these services of a recommended approach for measuring the net are precious to the nation. accumulation of timber. The section ends with the In the remaining sections we explore the issues panel’s conclusions on forest resources. raised in the preceding sections in far more detail for the cases of forests and air quality. TheNatureofForestValues

FORESTS: A RENEWABLE NATURAL Forests produce economic value through three RESOURCE principal classes of economic goods: private goods traded in markets, private goods not traded in mar- Forests are a prime example of renewable natural- kets, and public goods. These goods can affect 6 resource assets. They present many of the same both the national asset accounts and the NIPA. national economic accounting issues as other These three classes of forest goods and services are renewable natural-resource assets, such as agricul- discussed in decreasing order of availability of data tural land, fisheries, and coastal and freshwater and of accepted analysis required to include them resources. Many of the products derived from in the national economic accounts. natural-resource assets are included in the pro- Private, market-related activities. Some forest- duction accounts of the existing core NIPA. But based market-related activitiesarealreadyincluded these assets are not generally included in national in the national income accounts; examples are asset accounts, and the production accounts them- all forest products used in manufacturing (log- selves exclude any nonmarket goods and services ging, lumber production, the manufacture of derived from these natural-resource assets. Forests paper, wooden furniture, and musical instru- are a useful example because much effort has been ments). Some fuel wood production would fall devoted internationally to forest accounting. into this category; the part that flows through the While the NIPA as currently structured are not market economy would enter the accounts, while intended to include the full range of forest values, the part that is produced for own consumption regular reportsofeconomic activityasmeasured by would not. the NIPA are widely noted and interpreted as mea- The major issue in the current treatment of pri- suring important aspects of economic well-being. vate, marketed forest-based goods and services is It is logical to try to capture in these accounts thefailuretoaccountforchangesinthevalueofthe more of the important relationship between forests standing timber. Most of the conceptual problems and humans. Forests support human material and involved in doing so have been fully considered spiritual welfare in countless ways. They harbor and developed, as discussed below. Accounting for many important species of plants and animals. changes in the timber inventory would address one They form an aesthetically pleasing backdrop for of the major shortcomings of the existing forest recreation and for everyday life. They filter and accounts. regulate the flow of much of the U.S. water sup- Private goods not traded in markets. Forests ply. They have been a reservoir for land available produce many private goods and services that— for conversion to agriculture and other developed for reasons of custom, law, or economics—society activities. Wood is one of the world’s most im- has elected not to allocate through markets.7 For portant industrial raw materials and a ubiquitous example, the water flowing from forested water- source of energy. And worldwide, literally millions sheds has considerable economic value. Indeed, of indigenous people call forests home. the rationale for forest conservation in the late This section examines, in five parts, method- nineteenth century related primarily to protec- ological and practical issues that arise with regard tion of forested upland watersheds. Protection to including forests in national economic accounts. 6. The following discussion focuses primarily on issues pertinent to the It begins with a discussion of the nature of the United States. A significant issue in natural-resource accounting for many economics of forest values, providing a general developing countries is deforestation. For example, a major concern in the national accounts of developing countries such as Indonesia is that harvesting framework for assessing those values. The second of forests is contributingtorapid growthin current consumption at the expense subsection translates this general discussion into a of the stock of forest assets. In the late 1800s, the deforestation rate in the United States equaled or exceeded that found in many tropical countries today, more precise statement of how forest values might but deforestation is no longer significant on a national scale, and the general be incorporated in the U.S. economic accounts. trend since the 1950s has been a net growth in the forest stock of the United Given this context, the third subsection comments States. 7. Because of the decision not to use markets in allocating such resources, on BEA’s work to date and provides a brief dis- but typically to provide them through collective decisions, common usage sometimes refers to such goods and services as “public goods.” This report cussion of the extensive international literature on follows the conventional definitions of public and private goods discussed in forest accounting. This is followed by discussion the previous section. 40 March 2000     •

of navigation was the explicit constitutional ba- Incorporation of Forest Values in the sis for creation of the eastern national forests, National Economic Accounts8 and congressional agricultural interests concerned about irrigation provided the principal support for To be most useful, the economic accounts would withdrawing the national forests from the west- identify the separable contributions of forests to ern public-domain lands. A study by Bowes et al. the national economy. It is convenient to discuss (1984) of the Front Range of the Rockies around the problems involved in incorporating forest val- Denver and informal estimates for the Quabbin ues in the U.S. national economic accounts first Watershed servicing Boston demonstrate that in for the production accounts and then for the asset some locations, the value of the water produced accounts. from a forest may far exceed the value of the timber production. Changes in forest attributes Adjustments to Production Accounts can affect stream flow and therefore the value of water “produced.” Interestingly, Bowes et al. A full treatment of forests in the production ac- (1984) demonstrate that when water is valuable, counts would involve the following adjustments to it is optimal to keep timber stocks low to reduce national income and product. evapotranspiration and therefore increase runoff. Timber income. Sales of timber are already in- cluded, although some are recorded as part of personal income, some as part of manufacturing Public goods. Public goods are ones for which income, and some as part of government receipts. consumption byoneindividual doesnot reducethe The principal difficulty is ascribing these income amount available for others to consume. Forests streams to the forest sector; in this respect, the produce many public goods, including aestheti- issues are very similar to those encountered in cally pleasing landscapes, a carbon sink, and a the treatment of mineral incomes discussed in store of biological diversity. Given data on changes Chapter 3. Ordinary production costs associated in forest inventories, it may be possible to value with forest production activities are similarly cov- some of these services (e.g., the value of carbon ered by the current NIPA, but may not be easily sequestration), although the uncertainties of such associated with the forests themselves, rather than valuation should not be underestimated. In other forest-products manufacturing. Problems remain cases, the valuation problems go far beyond the with the allocation of joint costs. For example, results of current research. forest roads are a costly input to the production of many forest products, including timber, mi- nor forest products, and recreation. Yet standard accounting practices, especially for the national The interactions among these three sources forests, attribute the full cost of these roads to the of forest value—private marketed goods, private timber program. As currently constructed, the nonmarketed goods, and public goods—can be NIPA include the costs of road construction, but complex. For example, cutting trees leads to in- exclude the benefits produced by the road. creases in manufacturing activity. This in turn Near-market forest products. To the extent that might cause an increase in water yields and thereby near-market forest products, such as fuel wood, reduce the costs of industrial and household pro- berries, mushrooms, and Christmas trees, are pro- duction. It might also cause a shift of species duced by households but not purchased through diversity away from late-seral-stage organisms, markets, they would be included in the forest such as spotted owls, and toward early-seral-stage accounts. ones, such as elk. It would lead to an imme- Contributions to household production (e.g., diate release of carbon associated with logging recreation). The accounts would include the value and forest products manufacturing, but might of household production of activities such as hik- result in a long-term increase in carbon seques- ing, hunting, and fishing. However, if there tration with forest growth if the wood products is uncongested, open access to the forest-based were sequestered in long-lived furniture or houses. inputs needed for household production, the con- Given the site-specific nature of such production tribution of these inputs to household value on the relationships and the lack of current scientific un- margin is zero. Current practice often uses average derstanding of many of the underlying ecological rather than marginal values, so care must be taken, processes, there is currently an insufficient scien-

tific basis for specifying a full set of such linkages 8. The discussion in this section draws heavily on the recentcomprehensive in supplemental accounts. treatment of the subject by Vincent and Hartwick (1997).     March 2000 41 • particularly for open-access forests, to ensure con- stringent emission controls under the Kyoto Pro- sistent valuation in order to prevent overvaluation tocol of 1997. Third, the Forest Service presents of nonmarket activities. different types of estimates for the value of forest Environmental services used by other industries services, market-clearing prices being only one of (e.g., watershed protection, domestic/industrial wa- these.9 ter supply). Some of the impacts of forests are already included in the NIPA. For example, if Forests Asset Accounting forests moderate water flows and reduce the cost of agricultural production, this benefit is fully in- A key conceptual problem with the present NIPA corporated in the NIPA. Ascribing the benefit to is the lack of any accounting for changes in asset the forest sector, while a difficult task, would be values of U.S. forests. Accomplishing this task was required for a full accounting. part of the Phase II work outlined by BEA (see Chapter 2). We address this issue in some detail Public goods (e.g., carbon sequestration, bio- for two reasons. First, from a conceptual stand- diversity, species preservation). At present, the point, natural-resource assets should be treated only public goods that have been the subject of consistently with produced capital assets, adding widespread attempts at valuation are those asso- net accumulation or subtracting net decumulation ciated with carbon sequestration (Brown, 1996). from gross domestic product (GDP) to arrive at While quantitative data on carbon sequestration a measure of net national product (NNP) more are available, valuation is still highly uncertain. closely associated with a sustainable-income con- Moreover, because valuation of carbon sequestra- cept. Second, the capacity exists to rectify this tion is based on global benefits, the issue of how omission with respect to the value of forests that is such benefits would be incorporated in a single linked to marketed production. nation’s accounts is unresolved. While adjustments in an asset account are con- There are few comprehensive studies of the total ceptually similar to net investment of “made valueofforestproducts. Recentworkongoodsand assets,” for forests it is more precise to call the services produced on public lands managed by the change in asset values net accumulation to reflect U.S. Forest Service indicates that more forestland the fact that, even at constant prices, the asset value is due to recreational and wildlife services value of a forest can either increase or decrease. than to timber, mineral, and range goods (U.S. Most generally, net accumulation is defined as the Department of Agriculture Forest Service, 1995). change in an asset value from one period to the For example, of the estimated total $9 billion value next. Because asset values cannot generally be of forest goods and services in 1993 (valued at inferred, economists infer the value of the asset market prices), recreational and wildlife services from assumptions about timber markets. A full accounted for 80 percent, whereas the production analysis of this issue is presented in Appendix C. of minerals and timber and grazing range services Three major alternative approaches to accounting accounted for just 20 percent. for changes in asset values of forests are described While the above estimates illustrate the impor- below. tance of nonmarket production, they should be Hotelling model. The first approach is analogous interpreted with caution. First, they include only to the literature on nonrenewable resources dis- land managed by the U.S. Forest Service, which is cussed in Chapter 3. In a sense, this approach treats not representative of all forestland. By contrast, the exploitation of primary, old-growth forests as on private lands that are intensively managed for timber mining. Since it is generally uneconomic to timber production, much of the value is due to replace primary forests with forests of a similarly timber harvesting. Second, these estimates do not old age, this analogy is not as odd as it might ap- include all nonmarket values; for example, they pear. Under these circumstances, the change in the omit the potential value of carbon sequestration. value is the volume of the harvest times the differ- A recent estimate is that U.S. forests sequestered ence between the price and the marginal extraction 211 million metric tons of carbon in 1992 (Birdsey cost. This model of net accumulation is called the and Heath, 1995). At $10 per ton, a value consis- tent with the Intergovernmental Panel on Climate 9. USDA Forest Service(1995) also present estimatesbasedon feescollected (which show much lower value overall and relatively less for recreation and Change (IPCC) estimates of the marginal value of wildlife); willingness to pay, including consumer surplus (which show higher emission reductions (see Bruce et al., 1996), the overall values and greater importance for recreation and wildlife); and income generated, including that generated by downstream activities such as lodging annual value of carbon sequestration in all U.S. and equipment rentals related to forestland recreation (which show the highest overall value). From the perspective of comparability with the current national forests would be $2.1 billion; the numbers could economic accounts, the methods associated with the discussion in the text are be an order of magnitude larger if the U.S. adopted preferable to the other three methods. 42 March 2000     •

Hotelling model to emphasize the connection be- difference between price and marginal harvesting tween mining old growth that will not be replaced cost times growth minus harvesting (rather than and mining minerals that cannot be replaced. simply minus harvesting in the Hotelling model). Based on historical studies, this approach ap- By recognizing forest growth, such a formulation pears to be a reasonable approximation of em- improves on the ordinary Hotelling approach, but pirical trends in forest development (see Berck, still suffers the defects of (1) ignoring endogenous 1979; Lyon, 1981; Sedjo and Lyon, 1990; and Sedjo, price changes in the sector, and (2) characteriz- 1990). In the early stages of development, net ing the forest only by net growth and not its more growth of the forest is nil: photosynthesis just bal- complex underlying age-class structure. ances the death of plant tissues and entire trees. Managed second-growth forests. Economic the- Because growth is nil, any harvest at all exceeds the ory suggests that, once the transition between growth of the forest. Since the harvest is greater old- and second-growth forests is complete, tim- than the growth, the timber inventory declines. As ber prices will stabilize, and the economic return the inventory of old-growth timber declines, tim- to holding forests will arise solely from forest ber becomes more scarce, and timber prices rise. growth. Vincent (1997) has analyzed this case and In addition, harvesting costs increase as logging developed the appropriate measures of net accu- extends into increasingly remote sites. Prices rise mulation for optimally managed second-growth until the purposeful husbandry of second-growth forests. The appropriate estimate of the value of timber and the use of nonwood substitutes (stone, asset accumulation is more complicated here (see concrete, and steel for construction; fossil fuels, so- Appendix C for a full discussion). Accumula- lar energy, and conservation for energy) becomes tion depends on the forest age structure, discount economic. This analysis is broadly consistent with rate, timber-yield function, and economically op- the development of the forest sector in the United timal rotation age. While this approach improves States. Harvest exceeded growth until the 1950s. on both the Hotelling and transition approaches, Timber prices rose at a real rate of about 4.6 percent certain shortcomings remain. In particular, this per year between 1910 and World War II and 3.1 approach assumes that forest owners cut their trees percent per year from that period to the mid–1980s at the economically optimal time and that timber (Clawson, 1979; Sedjo, 1990; and Binkley and prices grow at a constant rate. This theory of for- Vincent, 1988). est valuation can be used to formulate a practical Transition models. While the Hotelling model approach to measuring the economic depreciation maybeappropriateforthecaseofpuredepre- of forests. Before turning to that recommended ciation under the assumption of perfect capital approach, it is useful to examine BEA’s work on markets,10 it misses several important aspects of forests and the international literature in this field. the forest sector, including (1) “discovery” of new old-growth forest stocks (e.g., the rapid expansion BEA’s Approach and International of logging in the British Columbia interior to serve Comparisons U.S.marketsonceU.S.priceshadrisentothepoint that accessing this comparatively remote region As noted, forests are part of Phase II of BEA’s IEESA became economic), and (2) the fact that the old- effort. As a consequence, BEA’s work on forests growth forests were replaced with faster-growing to date has not been extensive and may need re- second-growth forests. Both effects attenuate price finement (see Howell, 1996). In its current work, increases, causing the ordinary Hotelling model BEA separates forestland from the timber inven- to overstate forest depreciation. These effects are tory. “Forests and other wooded land” are valued the forest analog of mineral deposits analyzed in at the average value of agricultural land. In general, Chapter 3. edaphic and geomorphologic factors make forest- Transition models account in part for these land less valuable than agricultural lands, and the problems by recognizing that forest growth offsets rate of change in forestland prices is uncorrelated harvests. Assuming constant prices and a forest with the rate of change in farmland prices (see inventory recognized only by total net growth, this Washburn, 1990). BEA updated their estimates model suggests net accumulation is given by the of the timber inventory each period using sepa- rate Forest Service estimates in physical terms of 10. The Hotelling model assumes perfect capital markets in which the rate of growth and removals. Starting with physical in- return in the mining or old-forest sector equals the rate of return in alternative economic activities. In countries, especially developing countries, where both ventory estimates, BEA added physical estimates forest and mining activities earn disproportionally high returns because of special favors and licenses, the Hotelling model is not appropriate. It greatly of growth (additions) and removals (depletion) to overstates the true decline in the value of these stocks as they are mined. derive closing stocks. Each year’s closing stock es-     March 2000 43 • timate became the following year’s opening stocks criticisms of this approach and plans to use other (except in the Forest Service inventory years, when approaches in the future. BEA publishes a full ac- inventory estimates of standing timber were used). count for 1987, although it produces data on the Opening and closing stocks, additions, and deple- value of timber stocks for 1952–1992. Using BEA’s tions were then valued at the stumpage prices; the data, the net accumulation of timber in 1987 was difference between the opening stocks plus addi- $2.1 billion at 1987 prices and $47.0 billion if price tions less depletion and closing stocks, in monetary changes are included. terms, was placed in revaluations. While BEA’s methods can and should be refined as the environmental accounts are developed, they BEA uses the Hotelling model to value the tim- are consistent with current international practice. ber stock in each period. Timber is valued at the Table 4–4 provides a summary of 29 studies from national average stumpage rate, with species di- around the world that have attempted to extend the vided into two categories, softwood and hardwood. treatment of forests in national income and prod- When measured at a national level, marginal ex- uct accounts. Most of these efforts use variants traction costs are probably nonzero (production of the so-called “net price” approach (see equa- increases are accomplished by turning to increas- tions C.3 and C.4 in Appendix C). Many fail to ingly costly regions). There is some evidence that distinguish marginal and average extraction costs. extraction costs are constant within regions, how- Accounting for net timber accumulation is well es- ever (Adams, 1997). One conceptual flaw in BEA’s tablished in the international literature. None of current approach is that it measures the deprecia- the studies appears to use the third method de- tion of recreational land on the basis of the costs scribed in the previous subsection of a managed of repair and maintenance of federal government second-growth forest. expenditures for parks. The panel has noted in numerous places the flaw in this approach. Hav- ing accounted for one of the costs of providing A Recommended Approach for Measuring recreational services, BEA does not adjust national Net Accumulation of Timber income to reflect the benefits. BEA recognizes the The three alternative approaches to accounting for changes in asset values of forests discussed above TABLE 4–4 Summary of Forest Accounting Studies incorporate many restrictive assumptions. The

Valuation Method panel investigated other alternatives and identified one (developed by Vincent [1997]) that is similar Study Area Reference Net El Price Serafy NPV Other to the second-growth forests approach, but allows

Global World Bank (1997) T U for the possibility that forest managers may devi- Asia Vincent and Castaneda G (1996) ate from ideal wealth-maximizing behavior. This Australia I Young (1993) U approach is described in detail in Appendix C. A Australia II Skinner (1995), Joisce HUU (1996) review of available data indicates that the approach Austria Sekot et al. (1996) H U U Canada I Anielski (1992a, 1992b, T can be readily implemented for the United States 1994, 1996) Canada lI Statistics Canada (1997), HU using data maintained by the U.S. Forest Service. Baumgarten (1996) Chile Claude and Pizarro (n.d.) ? ? ? ? China Li (1993) T Costa Rica I Repetto et al. (1991) ? ? ? ? Conclusions on Forest Resources Costa Rica II Aguirre (1996) T Ecuador Kellenberg (1995) T U Finland I Koltolla and Mukkonen T BEA has initiated a useful effort to recognize the (1996) Finland II Hoffren (1996) T economic contributions of forests in the NIPA. Indonesia Repetto et al. (1989) T Malaysia I Vincent et al. (1993) T Doing so is consistent with a wide international Malaysia II Vincent (1997), Vincent et G interest in such accounts. The data and methods al. (1997) Mexico van Tongeren et al. (1993) T U employed by BEA to date are reasonably consis- Nepal Katila (1995) T New Guinea Bartelmus et al. (1992, XX XX tent with the body of international work in this 1993), Bartelmus (1994) New Zealand Bigsby (1995) H area. At the same time, data are available for U.S. Philippines I IRG et al. (1991, 1992) T U forestlands that can enable much more complete Philippines II Cruz and Repetto (1992) T Sweden I Hulkrantz (1992) T estimates of net timber accumulation than either Sweden II Eliasson (1996) T Tanzania Peskin (1989a) X X X those developed to date by BEA or those available Thailand Sadoff (1993, 1995) T U United States Howell (1996) H in the literature for other countries. BEA could Zimbabwe Crowards (1996) T fruitfullyworkwiththeU.S. Forest Servicein devel- Key: H = Hotelling approach; T = transition approach; G = generalized El Serafy approach oping annual estimates of net timber accumulation (elasticity of marginal cost not infinity); X = no timber valuation performed; ? = no information; U = used technique; NPV = net present value. using these data. Source: Vincent and Hartwick (1997). References in original. 44 March 2000     •

This work could also be related to other im- secondreviewssomemajorpollutantsthatresultin portant values of the forest, particularly recreation degradation of air quality and their primary phys- and other nonmarket activities. While the data ical effects. This is followed by review of a recent and analytical methods are not yet adequate to attempt to estimate comprehensively the benefits provide precise estimates of the value of all forest- associated with improvements in air quality. The sector flows to the economy, nonmarket forest fourth subsection addresses the relevance of these values for the nation as a whole appear to ex- damage estimates to environmental accounting. ceed the value of timber by a substantial amount. The section ends with the panel’s conclusions on Many of these forest values (such as recreation accounting for air quality. or self-produced fuel wood) are best understood conceptually in the context of household produc- tion. The household combines specific aspects Air Quality Impacts on Market and of the forest resource with household capital and Nonmarket Activities labor to produce valuable nonmarket goods and services. Viewed in this context, forests present Degraded air quality can have a harmful effect on many of the same challenges for national account- both market activities (e.g., reduced crop yields ing as do such important products and services or lost work-days) and nonmarket activities (e.g., as home-cooked meals and in-home education or losses due to illness beyond those related to paid childcare. It is therefore logical for BEA to consider labor, such as those to retired persons, and reduced these aspects of environmental accounting as part amenities in recreational facilities). These air qual- of the larger problem of valuing the contributions ity effects belong in the production accounts of of nonmarket activity to economic well-being. environmental accounts. Moreover, degraded air In conclusion, constructing a set of forest quality can affect the value of natural-resource as- accounts is a natural next step in developing inte- sets (e.g., acid deposition damage to forests), can grated economic and environmental accounts. At cause deterioration of physical capital (e.g., dam- the same time, it must be recognized that there age to the exterior of buildings), and has long-term are many thorny problems involved in forest ac- health impacts that affect human capital (e.g., pre- counting. Given the available data and methods, mature death and effects of lead on measured IQ the panel concludes that this accounting is a useful of children). Such effects might be included in the next step in developing the IEESA. asset component of environmental accounts. With assets as with production, there are both market and nonmarket effects: market impacts include AIR QUALITY: capital asset deterioration and forest timber loss, A PUBLIC ENVIRONMENTAL GOOD while nonmarket impacts include lost value due to damaged landmarks or degradation of forests for Air quality is one of the most important exam- recreational purposes. ples of a public environmental good and thus should be among the top priorities for inclusion in environmental accounts. It also presents issues Major Air Pollutants and Their Health and for environmental accounting similar to those en- Ecological Effects countered with other environmental assets, such as water quality and climate change. Severely de- Table 4–5 lists some important health and ecologi- graded air quality in many cities of the United caleffectsofexposuretosixairpollutantsforwhich States in the 1960s generated a number of fed- the U.S. Environmental Protection Agency (EPA) eral regulations during the early 1970s designed has established National Air Quality Standards— to reduce emissions of pollutants that contributed carbon monoxide,ground-level ozone,lead, nitro- to this degradation. Air quality has many di- gen dioxide, particulate matter, and sulfur dioxide. mensions, and early regulations focused on some These chemicals are sometimes referred to as “cri- of the more obvious and easily addressed prob- teria pollutants.” In addition, there are many other lems. As scientific research further illuminated the constituents of the atmosphere that may have im- less immediately obvious impacts of degraded air pacts of economic consequence. Table 4–6 lists quality, such as chronic effects on health, these ear- some other components of air pollutants, includ- lier controls were tightened, and new regulations ing air toxins (e.g., benzene), stratospheric ozone addressed a wider range of pollutants. depletors (e.g., CFCs), and greenhouse gases (e.g., The first subsection below examines the various carbon dioxide and methane). As indicated, EPA market and nonmarket impacts of air quality. The has identified 188 air toxins alone.     March 2000 45 •

Exposure to air pollution has a wide range of im- tants from 1986 to 1995.11 Primarily as a result of pacts, including respiratory illnesses (which result the Clean Air Act and the Clean Air Act Amend- from ground-level ozone, sulfur dioxide, nitrogen ments, emissions of the six primary pollutants have dioxide, particulate matter, and air toxins); child decreased substantially. For example, installing IQ loss, infant mortality, strokes, and heart attacks scrubbers and switching to low-sulfur coal caused (which result from lead); skin cancer (which is the a 19 percent decline in emissions from coal utility indirect consequence of stratospheric ozone de- plants, which in turn resulted in an overall 18 per- pletors); and increased mortality (resulting from cent decline in sulfur dioxide emissions from 1986 particulate matter, lead, and air toxins) (see Pearce to 1995. A 16 percent decline in carbon monoxide et al., 1996). Ecological effects include impacts emissions during the same period resulted primar- on agricultural, forest, and aquatic ecosystems. ily from a 20 percent decline in carbon monoxide Airborne chemicals have both positive and nega- emissions of on-road motor vehicles. Similarly, a tive effects on production of marketed goods and 32 percent decline in lead emissions was primarily services. Ground-level ozone harms crops, while a result of the ban on leaded gasoline. nitrogen deposition and carbon dioxide enhance Declines in nitrogen dioxide (14 percent) and plant and timber growth. Ground-level ozone ground-level ozone emissions (6 percent) were less and sulfur dioxide reduce crop yields and tim- dramatic, but are expected to become more pro- ber growth, while air toxins and sulfur dioxide nounced as the Clean Air Act Amendments of 1990 reduce freshwater fish yields. In other cases, at- become effective. For example, reformulated fuel mospheric trace gases have subtle effects that will requirements (for oxygen and volatility) for on- occur far in the future affecting biological diver- road vehicles are likely to reduce carbon monoxide sity (for greenhouse gases) or ocean food web and ground-level ozone emissions. Similarly, the stresses, and ultimately causing severe sight dam- Acid Rain Program (Title IV) requires a 40 percent age for many mammals (for stratospheric ozone reduction in sulfur dioxide and a 10 percent re- depletors). duction in nitrogen dioxide emissions from 1980 Table 4–5 also shows the change in emissions and to 2010. Particulate matter may be more diffi- sampled concentrations of EPA’s six criteria pollu- cult to control given that almost 70 percent of anthropogenic-related emissions result from fugi- tive dust (e.g., unpaved roads), with an additional TABLE 4–5 Environmental Protection Agency’s Six Criteria 20 percent coming from agriculture and forestry. Air Pollutants Thedeclinesinemissionsare,ofcourse,linkedto Pollutant Trends (1986–1995) Major Effects Leading Source lower concentrations of the six primary pollutants. Whereas emissions are estimated on the basis of Ground-level ozone (O 3) Respiratory illness/lung Transportation* (37%) Concentration –6% damage Solvent utilization (28%) Emissions –9% Crop/forest damage Building/material damage 11. Data prior to 1986 exist, but cannot be directly compared with data Visibility problems collected from 1986 on because of changes in data collection (see U.S. Environmental Protection Agency, 1996, for more details). Carbon monoxide (CO) Reduced oxygenation of Transportation (81%) Concentration –37% blood Emissions –16% Heart damage TABLE 4–6 Other Pollutants of Air Quality Identified by Sulfur dioxide (SO 2) Respiratory illness Electric utilities (66%) Concentration –37% Building/material damage Environmental Protection Agency Emissions –18% (acid rain) Crop/forest damage Pollutant Major Effects Leading Source Visibility problems Air toxins (188 in total, Thought to cause cancer Transportation, wood Nitrogen dioxide (NO 2) Respiratory illness/lung Transportation (49%) Concentration –14% damage Electric utilities (29%) e.g., dioxins, benzene, or other serious health combustion, chemical Emissions –3% Building/material damage arsenic, beryllium, effects, such as birth plants, oil refineries, (acid rain) mercury, vinyl chloride) defects or reproductive aerospace, manufac- Crop/forest damage effects tures, dry cleaners Visibility problems Ecosystem damage (par- ticularly freshwater Lead (Pb) Infant mortality Metals processing (smelt- fish) Concentration –78% Reduced birth weight ers, battery plants) Emissions –32% Childhood IQ loss (39%) Stratospheric ozone Skin cancer Fossil fuel, industrial Hypertension Transportation (31%) depleters (e.g., Cataracts cleaners Heart attacks chlorofluorocarbons Suppression of the im- [CFCs], halons, carbon mune system Particulate matter (PM- Lung disease Fugitive dust (68%) tetrachloride, methyl Ocean food chain 10) Mortality Agriculture and forestry chloroform) stresses Concentration –22% (20%) Emissions –17% Greenhouse gases (e.g., Broad-scale changes in Fossil fuel, combustion, carbon dioxide, temperature and pre- landfills * Based on volatile organic compounds (VOC) emissions. methane, halogenated cipitation affecting agri- fluorocarbons [HFCs]) culture, health, water Source: U.S. Environmental Protection Agency (1996). resources, recreation, ecosystems Sea level rise

Source: U.S. Environmental Protection Agency (1996). 46 March 2000     •

industrial activity, technology, fuel consumption, of actual air pollution relative to a counterfactual and vehicle miles traveled, concentrations of pol- baseline that assumes no controls imposed after lutants are measured at selected monitoring sites 1970; roughly speaking, the counterfactual is for across the country. Based on these measurements, emissions to grow with the economy, rather than estimated airborne concentrations of lead have declining as described above. The major result fallen by 78 percent since 1986, while concentra- presented is that the economic benefits of reduced tions of airborne carbon monoxide,sulfur dioxide, air pollution in 1990 are estimated to be worth and particulate matter have fallen by 37, 37, and 22 $1,248 billion. Reduced mortality benefits ($1,004 percent, respectively. Smaller declines occurred for billion) account for 80 percent of this total; to- ground-level ozoneand nitrogen dioxide (6 and 14 gether, avoided human health effects account for percent, respectively). 99 percent of the total. In addition, benefits of Data on other air chemicals vary widely. Excel- improved visibility are estimated at $3.4 billion, lent data are available on emissions and concentra- those of reduced household soiling at $4.0 billion, tions of many of the greenhouse gases (particularly and those of increased agricultural income from carbon dioxide) and stratospheric ozone destroy- reduced yield losses due to ozone at about $1.0 ers. EPA presently monitors national ambient billion. With regard to specific pollutants, most concentrations for few of the 188 air toxins identi- of the benefits are attributed to reductions in par- fied in the Clean Air Act Amendments. Rather, the ticulate matter (PM–10) and lead; the benefits of agency sets technology-based performance stan- ozone reduction are estimated to be only on the dards to control emissions of these substances. As order of $2 billion. a result, EPA has only begun developing a National Caution is warranted in drawing too many con- Toxins Inventory. clusions from these estimates and comparisons. Certain assumptions might have had the effect of exaggerating the economic benefits, and there are Monetized Benefits of Clean Air Regulations major uncertainties about the health impacts, par- ticularly because of weaknesses in human exposure Although a great deal of work has been done on data. Moreover, the study omits some of the ma- valuing components of air quality, there is cur- jor effects of acid deposition on forests, lakes, and rently no comprehensive measure of the economic buildings, and the impact of tropospheric ozone impacts of air pollution for the United States. on ecosystems is not valued. The figures presented However, a recent EPA study evaluating the eco- should therefore be viewed as order-of-magnitude nomic costs and benefits of clean air regulations estimates. Even with all these qualifications, how- provides a useful benchmark that sheds light on ever, it appears that the economic impacts of air this issue (U.S. Environmental Protection Agency, qualityonhumanhealtharehighlysignificant. 1997). The estimates given are subject to many uncertainties due to the difficulty of estimating exposure and the incidence of effects related to ex- Air Quality Benefits and Supplemental posure and valuing the effects. In addition, data Accounts on air toxins have only recently become available, making it difficult to develop comparableestimates The estimates of the benefits of pollution con- for these pollutants. The EPA study includes no trol just discussed reflect the value of changes in physical or monetary assessments of the impacts of the level of air pollutants resulting from proposed changes in air quality on ecosystem health, physi- regulations. They are relevant for regulatory or cal capital, or global public goods, such as slowing cost-benefit purposes, but they are not the appro- climate change and preventing ozone depletion. priate values for economic accounts. Production Moreover, many of the estimates of benefits, par- accounts should measure the damages associated ticularly those involving the valuation of health with remaining levels of pollution, in terms of both benefits and the discount rate, have been the sub- production accounts and change in asset values. ject of major criticism (see Clean Air Act Council This difference between abatement and residual on Compliance, 1997). damage can be quantitatively large. For example, Notwithstanding these limitations, the EPA ozone concentrations fell only 6 percent between study provides an indication of the overall eco- 1986 and 1995. As a result, regardless of the bene- nomic importance of changes in air quality, as well fits of preventing higher levels of ozone than those as a sense of the relative importance of the var- of 1986, the value of changes in ozone concentra- ious air pollutants and the impacts on different tions over this period would be relatively small. In sectors. The study estimates the economic benefit contrast, lead and PM–10 concentrations fell 78     March 2000 47 • and 22 percent, respectively, over the same period, proved sanitation, vaccinations, and public-health and consequently the damages from these chemi- measures—led to improved life expectancy over cals would be much smaller in 1995 than in 1986. the first seven decades of this century. It would In other words, whereas comprehensive consump- therefore be misleading to enter only a large health tion would have a substantial negative entry due to negative into a set of augmented income accounts. lead and PM–10 in 1986, the negative values would The positives and negatives in the environmental be of much smaller magnitude in 1995. The result entry in a set of health accounts would have to be might be a substantial increase in the estimate of placed in the context of the vast changes in health growth of comprehensive consumption over this status of the American population. period. As discussed earlier, air pollution affects pro- duction activities, assets, and nonmarket activities. Conclusions on Air Quality Most of the estimates from the EPA study refer to the production accounts: days of work lost, The basic finding emerging from the above dis- shortness of breath and acute bronchitis, loss of cussion is that air quality is likely to be a major visibility, and crop losses are effects on produc- nonmarket effect. WhileEPA’sestimatesofbenefits tion activities. Crop losses and the output losses of $1.2 trillion per year due to reduced air pollu- from lost work-days are already included implic- tion are highly uncertain, do not include all effects, itly in the accounts because these relate to market and measure a somewhat different concept than activities. Supplemental accounts that would iden- would be appropriate for the accounts, it is likely tify these losses separately would serve to connect that a realistic assessment of reduced damages due them specifically to air pollution. The estimates for to improved air quality would yield a much larger shortness of breath and acute bronchitis include figure than the $27.1 billion in air pollution control both damages that may already be reflected in the expenditures used by BEA as a placeholder. In the production accounts (i.e., reduced worker produc- panel’s view, no other area of natural-resource and tivity while on the job) and damages that would environmental accounting would have as great an be reflected only if the accounts were expanded impact as the potential correction from air qual- to include household production (e.g., impacts on ity. The magnitude of this impact indicates that tennis and jogging). Many of the effects not es- the development of supplemental accounts for air timated by EPA, such as those of acid deposition quality is a high priority. Indeed, the overall review on forest health, freshwater quality, or ecosystem of augmented accounting in Chapter 2 reveals only function, would also include effects on both mar- a few areas close in importance, such as the value of ket activities already in the accounts, such as timber leisure, health status, and nonmarket educational or commercial fishing, and nonmarket goods, such investments. as recreation. At the same time, air quality is a most elusive Asset effects present greater complexity, as was concept since it has so many different compo- seen above for the case of forests. Some impacts, nents. To include these effects in the accounts, such as those on soil or fish farms, would be re- several data and measurement obstacles must be flected in the market value of these assets. Others, overcome. First, determination of the physical such as mortality and chronic bronchitis, are long- impacts of changes in air quality, generally esti- term effects on human resources. These effects mated through dose-response functions, should be would require adjustments in the asset accounts if focused on the effects of actual human exposure a full set of asset accounts for human health and to air pollution. Second, the damage estimates capital were constructed. must separate the market effects of changes in air One particular concern arises if the accounts are quality that are currently captured in the accounts to include the impact of air pollution on human (lost productivity) from the nonmarket effects that health. The impact of air pollution and other en- are not currently captured (lost leisure activities). vironmental activities on human health is often Third, there is a need for reliable and objective taken out of the context of other health-related ac- physical and monetary damage estimates associ- tivities. If one were to track environmental trends ated with exposure to air pollutants, including alone, it might be concluded that until the 1970s, air toxins, ozone depletors, and greenhouse gases. growing environmental problems were leading to Fourth, significant data gaps with respect to the a deterioration in the health status of Americans. impacts of air pollution and changes in air quality This conclusion is, in fact, incorrect. Activities on ecosystem health must be filled. And finally, outside the environmental arena—including im- the estimates must represent year-to-year changes, 48 March 2000     •

rather than changes from a hypothetical level of values that are based on actual behavior. Contin- pollution without regulations. gent valuation, while sometimes useful for other Developing a set of accounts in this area, along purposes, is currently of limited value for environ- with the associated physical measures and val- mental accounting in the context of the economic uations to apply to those measures, is a major accounts. long-run task for the nation. This task far tran- Valuing environmental goods and services re- scends the scope and budget of BEA, and much of quires distinguishing between private and public the necessary work lies outside BEA’s specialized goods. Market prices provide the marginal valua- expertise. The task for the short run, therefore, is tions for private goods, but determining the value to continue basic research on the underlying sci- of public goods requires the summation of individ- ence and economics of estimating the benefits of ual values. Moreover, there may be no behavioral public goods such as clean air. Many years of con- traces for individual valuation of public goods. certed research are likely to be required before the Price data are relatively reliable for private mar- materials for a set of augmented accounts in this ketgoodsproducedfromforestandagricultural area are available. But the payoff from the research assets, such as timber stumpage, livestock, and would be large, not only in producing the raw ma- land use and quality. Values for near-market terials for improved environmental accounts, but goods—those that have direct counterparts in the more important in providing the data and analysis market—can be constructed by comparing the needed for improved public policy concerning the near-market goodswiththeir market counterparts, environment. In short, the task of constructing en- adjusting for quality as necessary. Techniques for vironmental accounts for important public goods valuation of public goods are still under devel- should be part of a more general goal of improving opment. Some techniques—such as hedonic or the nation’s information and analytical systems in travel-cost studies—rely on behavioral or market- this area. based estimates; whiletheseestimatesare subject to significant measurement errors, they are concep- CONCLUSIONS AND tually appropriate in economic accounts. Other RECOMMENDATIONS ON RENEWABLE techniques, such as contingent valuation, are not AND ENVIRONMENTAL RESOURCES based on actual behavior, are highly controversial, and are subject to potential response errors. General Approach Quantitative Data 4.1 The panel recommends that BEA continue its work toward accounting for changes in natural- 4.3 Quantitative data on many natural-resource as- resource assets and forthe flow of services from these sets are currently relatively adequate. However, assets. the data on many environmental variables are at Environmental variables affect economic well- present poorly designed for the construction of en- being in three major ways: direct effects on con- vironmental accounts. The panel recommends that sumption or income of households, industry, and greater emphasis be placed on measuring effects as government; accumulation in the environment of directly as possible. Of particular importance are stocks of residuals that then affect economic activ- measuresof actual human exposure to air and water ities or economic assets; and effects on the service pollutants, rather than modeled measures of expo- flows of economic assets, including capital stock, sure based on ambient pollutant levels at current natural resources, and human resources. The main monitoring sites. value of natural-resource accounting is in provid- Quantitative data for natural resources are often ing a complete picture of the role these resources of high quality relative to the other quantitative play in the economy. Sometimes this information data in the NIPA because there are well-established can be used to judge the overall sustainability of the units of measure for many natural resources. use of resources, while at other times it can be used Quantitative data on near-market activities such as to manage natural and environmental resources fuel wood for own use are conceptually straight- and to inform public policy choices. forward, and many of these data are currently collected by federal agencies. Measurement of Valuation nonmarket goods and services and explicit ac- counting for quality changes, particularlyfor those 4.2 For valuation, the panel recommends that BEA that have public-good characteristics, are currently rely primarilyon market values or proxiesof market subject to severe methodological difficulties and     March 2000 49 • insufficient data. There are relatively good data flows may not be important to the national econ- on emissions of many residuals from industrial omy, they could be far more important to regional and human activities, but for most harmful pol- and local economies. lutants except lead there is very little systematic monitoring of human exposures. Next Steps

4.7 The panel recommends that funds be provided Inclusion of Public Goods to reinitiate and improve the design of the collec- 4.4 The panel finds that more work will be needed tion of data on pollution control and abatement on techniques for establishing production flows and expenditures. values for the assets and services of public goods to 4.8 As BEA further develops its natural-resource place them on a comparable basis with the prices and and environmentalaccounts, an important step is to quantities used in the core accounts. incorporate near-market goods and services—those True public goods, for example biodiversity, that have close counterparts in marketed goods and species preservation, and national treasures such services. There is a clear basis here for measur- as the Florida Everglades and Yellowstone National ing quantities and establishing values in a manner Park, present severe conceptual and measurement comparable to that used for the core accounts. issues for incorporation into a national accounting 4.9 Construction of a set of forest accounts is system. a natural step in developing integrated economic- environmental accounts. The United States has much of the data needed for such an effort, and the Data Collection analytical techniques are relatively well developed. 4.10 Based on available information, theeconomic 4.5 The panel encourages BEA to help mount a impacts of air quality are likely to be the most concerted federal effort to identify the data needed significant element in the environmental accounts; for measuring changes in the quantity and quality development of such accounts is a centraltask for en- of natural-resource and environmental assets and vironmental accounting. At the same time, because associated nonmarket service flows. of the unresolved conceptual issues and the need for Many different federal agencies collect data or appropriate physical measures, the development of have expertise that will be essential to BEA, particu- stock and flow accounts for air quality and other larly as its efforts expand to include Phase III assets important public goods poses awesome difficulties. and associated flows. BEA already cooperates with This task far transcends the scope, budget, and ex- other agencies in collecting data for the core ac- pertise of BEA. A major goal for the near term is counts; supplemental environmental accounts will to continue basic researchon the underlying science require cooperation with, for example, the Envi- and economics in this area. ronmental Protection Agency, the Department of Agriculture, the Department of the Interior, the References Bureau of Labor Statistics, the Bureau of the Cen- sus, the Energy Information Administration, the Anderson, R., and M. Rockel. 1991. Economic National Institute of Environmental Health Sci- Valuation of Wetlands. Discussion Paper #65, ences, and the Department of Health and Human American Petroleum Institute. Washington, DC. Services. Arrow,K.,R.Solow,P.Portney,E.Leamer, R. Radner, and H. Schuman. 1993. Report Regional Resolution of the National Oceanographic and Atmospheric Administration Panel on Contingent Valuation. 4.6 The panel recommends BEA focus on developing Federal Register (January) 58:4601–4614. supplemental accounts for the nation as a whole as Berck, P. 1979. The economics of timber: A a first priority. At the same time, BEA should pre- renewable resource in the long run. Bell Journal of serve regionaldetail whereit exists so that these data Economics 10:447–462. are available for analysts interested in developing Binkley, C.S., and J.R. Vincent. 1988. Timber accounts at the regional level. prices in the U.S. South: Past trends and outlook The development of national estimates will for the future. Southern Journal of Applied Forestry require sampling, measurement, and valuation 12:15–18. techniques that reflect the fact that the quality Birdsey; R.A., and L.S. Heath. 1995. Carbon and value of natural-resource assets and associated changes in U.S. forests. 70 pp. in L. Joyce, ed., Pro- flows vary geographically. While some assets and ductivity of American Forests and Climate Change. 50 March 2000     •

General Technical Report RM–271, USDA FS Pub., Howell, S.L. 1996. A review of the concep- Rocky Mountain Forest and Range Experimental tual and methodological issues in accounting for Station, Fort Collins, CO. forests. In Proceedings of Third Meeting of the Lon- Bowes, M., J. Krutilla, and T. Stockton. 1984. don Group on Natural Resource and Environmental Forest management for increased timber and water Accounting. Stockholm: Statistics Sweden. yields. Water Resources Research 20:655–663. Lyon, K.S. 1981. Mining of the forest and Braden, J.B., and C.D. Kolstad, eds. 1991. the time path of the price of timber. Jour- Measuring the Demand for Environmental Quality. nal of Environmental Economics and Management Amsterdam: North-Holland, Elsevier Publishers 8:330–344. B.V. Mitchell, R.C., and R.T. Carson. 1989. Using Brown, S. 1996. Managing forests for mitigation Surveys To Value Public Goods: The Contingent Val- of greenhouse gases. Pp. 773–778 in R.T. Wat- uation Method. Washington, D.C.: Resources for son, M.C. Zinyowera, and R.M. Moss, eds., Climate the Future. Change 1995: Impacts, Adaptations, and Mitigation Pearce, D.W., W.R. Cline, A.N. Achanta, S. of Climate: Scientific and Technical Analyses.Pub- Fankhauser, R.K. Pachauri, R.S.J. Tol, and P. Vel- lished for the Intergovernmental Panel on Climate linga. 1996. The social costs of climate change: Change. Cambridge: Cambridge University Press. Greenhouse damage and the benefits of control. Bruce, J.P., H. Lee, and E.H. Haites, eds. 1996. Pp. 145–178 in J.P. Bruce, H. Lee, and E.H. Haites, Climate Change 1995: Economic and Social Dimen- eds., Climate Change 1995: Economic and Social Di- sions. Published for the Intergovernmental Panel mensions. Cambridge, U.K: Cambridge University on Climate Change. Cambridge, U.K.: Cambridge Press for the Intergovernmental Panel on Climate University Press. Change. Bureau of Economic Analysis. 1994a. Integrated Portney, P.R. 1994. The contingent valuation economic and environmental satellite accounts. debate: Why economists should care. Journal of SURVEY OF CURRENT BUSINESS April: 33–49. Economic Perspectives 8(4):3–17. Bureau of Economic Analysis. 1994b. Ac- Randall, A., and J. Stoll. 1983. Existence value counting for mineral resources. SURVEY OF CURRENT in a total valuation framework. In R. Rowe and L. BUSINESS April: 50–72. Chestnut, eds., Clawson, M. 1979. Forests in the long sweep of Managing Air Quality and Scenic Resources at American history. Science 204:1168–1174. National Parks and Wilderness Areas.Boulder,CO: Clean Air Act Council on Compliance. 1997. Westview Press. Report of the Clean Air Act Council on Com- Samuelson, P.A. 1954. The pure theory of pub- pliance. Letter report to the EPA Administrator, lic goods. The Review of Economics and Statistics private communication. 36(4):387–389. Cornes, R., and T. Sandler. 1986. The Theory of Samuelson, P.A. 1955. Diagrammatic exposition Externalities, Public Goods, and Club Goods. New of a theory of public expenditure. The Review of York: Cambridge University Press. Economics and Statistics 37(4):350–356. Costanza,R.,R.d’Arge,R.deGroot,F.Farber, Sedjo, R. 1990. The Nation’s Forest Resources. M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.V. Discussion Paper ENR 90–07. Washington, DC: O’Neill, J. Pervello, R.Q. Raskin, P. Sutton, and Resources for the Future. M. Van den Belt. 1997. The value of the world’s Sedjo, R.A., andK.S. Lyon. 1990. The Long-Term ecosystem services and natural capital. Nature Adequacy of World Timber Supply. Washington, (May) 387:253–260. DC: Johns Hopkins University Press. Diamond, P.A., and J.A. Hausman. 1994. Con- Smith, V.K. 1993. Nonmarket valuation of en- tingent valuation: Is some number better than vironmental resources: An interpretive appraisal. no number? Journal of Economic Perspectives Land Economics 69(1):1–26. 8(4):45–64. U.S. Department of Agriculture Forest Service. Green, I.M., C. Folke, K. Turner, and I. Mate- 1995. Chapter 4, Socioeconomic Effects and Im- man. 1994. Primary and secondary valuation of plications of the Proposed Program, 1995 draft wetland ecosystems. Environmental and Resource RPA program. Economics 4:55–74. . Hanemann, W.M. 1994. Valuing the environ- U.S. Environmental Protection Agency. 1996. ment through contingent valuation. Journal of National Air Quality and Emissions Trend Report, Economic Perspectives 8(4):19–43. 1995. EPA 454–R–96–015 Office of Air Quality     March 2000 51 •

Planning and Standards, Research Triangle Park, tute for International Development, Cambridge, NC, October. MA. U.S. Environmental Protection Agency. 1997. Vincent, J., and J.M. Hartwick. 1997. Account- The Benefits and Costs of the Clean Air Act, 1970 ing for the Benefits of Forest Resources: Concepts to 1990. Draft, April. Office of Air and Radia- and Experience. Draft, July 10. FAO Forestry tion/Office of Policy Analysis and Review/Office of Department. Policy, Planning, and Evaluation. Washburn, C.L. 1990. The Determinants of For- Vincent, J. 1997. Net Accumulation of Tim- est Value in the U.S. South. Ph.D. thesis, Yale ber Resources. Manuscript, Harvard Insti- University.